The Night You Should Have Closed Early
Two guests, three staff, two hours left. Nobody has the authority to end it, so it runs to the scheduled time. What staying open actually costs, and how to make closing early a decision rather than a failure.
What staying open costs
Wages against nothing. The visible part and the smaller one.
Tomorrow's energy. ⚠️ Two hours of standing about is more tiring than two busy hours, and it is taken from the next shift.
Morale, specifically. People know when their time is being spent for no reason. It reads as not being trusted with an obvious judgement.
The close itself gets worse. ⚠️ A clean-down at 03:00 after two dead hours is worse than one at 01:30, so tomorrow starts from a worse room.
What closing early costs
Real, and it should be counted honestly:
The people who would have come. ⚠️ Rare at that hour and not zero — and the ones you turn away at the door remember it.
Reliability. A venue that might be closed is one people stop trying. This is the serious cost and it is cumulative.
The regular who arrives at 01:30. They may be your best customer, and their evening is the one you ended.
Making it a decision
Someone must have the authority. ⚠️ Named, in advance. Without it the default is always "stay open", regardless of the arithmetic.
Set the conditions. "Below N guests at 01:00, and no bookings, the manager may close." A stated line means nobody has to justify a judgement.
Decide what "closed" means. Doors locked, or serving the people who are in and not admitting more? ⚠️ The second is usually right and it is rarely specified.
Never close on a guest who is still spending. Obvious, and worth writing down.
The reliability problem
The strongest argument against early closing, and it has an answer.
Publish honest hours. ⚠️ If you never trade past 01:00 on a Tuesday, close at 01:00 on Tuesdays — reliably. That is better than nominally being open until 03:00 and sometimes not.
Reliable shorter hours beat unreliable longer ones, and this is the change most venues should make instead of deciding it night by night.
Look at the pattern first
Before deciding anything on the night, look at the month.
Takings by hour, by night of week. ⚠️ The last hour is frequently a small fraction of the night's takings, and the count is the whole argument.
If the same two hours are consistently empty, that is an opening-hours decision, not a closing-early decision — and it should be made in daylight, once, rather than at 01:00 by whoever is tired.
The number that makes it a decision
Marginal cost of the last hour staff on shift + power + one manager
Marginal revenue of the last hour covers × spend per head, that hour, by weekday
⚠️ Both are knowable in a week. Most venues argue about this from memory instead.
Where the cost hides
The hour itself obvious
The close that follows it 30-45 minutes, always
Tomorrow's late start sometimes
⇒ ⚠️ The last hour is never one hour. Price it as ninety minutes and the arithmetic changes.
Decide it in advance, not on the night
On the night whoever is tired decides
In advance the same rule applies every night
⚠️ A rule decided at 1am is a mood. Write the trigger while the room is quiet and everyone is rested.
A trigger that works
"If covers are below N at [time], last orders is called."
⇒ ⚠️ Name the number and the time. "If it's dead" is not a trigger; it is a feeling, and it will be applied inconsistently.
What closing early actually costs
The regulars who came late the real cost
Staff hours saved the visible benefit
The message it sends the one nobody prices
⚠️ Three is the expensive one. A venue that might be shut is a venue people stop trying — and you never see the visits that did not happen.
The mitigation
Publish it. "Sunday and Monday we close at midnight."
⇒ ⚠️ A published early close costs almost nothing. An unpredictable one costs the week.
Before cutting the hour, try cutting the cost
1. One fewer person on the last hour
2. Consolidate to one section
3. Start the close during the hour, not after it
⚠️ Three recovers most of the saving without losing a single cover.
Read the pattern by weekday, not by night
One quiet Tuesday noise
Four quiet Tuesdays a pattern you can act on
A quiet Saturday something else happened; find out what
⚠️ Never change hours on one night's evidence. The variance between two of the same weekday is larger than most of the effects you are chasing.
It is 01:00. Two guests, neither buying. Three staff. Two hours until close.
Everyone knows how this ends. Nobody has been told they may end it.
Three to hold
Takings in the final two hours, by night. ⚠️ The number nobody has produced, and it usually settles the question immediately.
Nights closed early, and what they took after the decision point. Check whether the judgement was right.
Guests turned away after an early close. Should be near zero. If it is not, the condition is set wrong.
Where the record has to sit
Answering "what does the last hour actually take" needs takings by hour, over months. If the record is a nightly total, the question cannot be asked and the hours stay as they were set years ago.
tasteck keeps takings against the orders and times that produced them, so the shape of a night can be read rather than estimated.
The two hours were not a judgement call. Nobody was allowed to make one, so the schedule made it — and the schedule was written before any of this was known.
Read next
Related articles
Who Types The Request In
A guest asks for somebody by name. Between that sentence and the figure on your monthly sheet, three or four people handle it. Every one of them is a place where it can quietly vanish.
What One More Regular Is Worth
Venues in this trade know what a quiet Tuesday costs them. Far fewer can say what one additional regular is worth over a year — which is the only number that tells you what you can sensibly pay to find one.
What a Listing Is Worth To One Room
A single-room salon and a twelve-room clinic get the same price from the same portal. Only one of them can work out whether that price is defensible, and it is the small one — because the arithmetic is smaller.
Map out your operations in 5 minutes
Eight questions cover reservations, customer management, shifts, and settlement. Results shown instantly with industry benchmark. Sales emails only if you request them.
Your answers are not stored. The assessment runs entirely in your browser.
Try tasteck free for 30 days
No credit card required. Full access to reservations, cast shifts, dispatch, and analytics.
- No card required
- Free data migration support
- All features unlocked for 30 days